Multi-Rate Deep Learning for Temporal Recommendation

Multi-Rate Deep Learning for Temporal Recommendation Yang Song, Ali Elkahky, and Xiaodong HeJuly 2016 Abstract Modeling temporal behavior in recommendation systems is an important and challenging problem. Its challenges come from the fact that temporal modeling increases the cost of parameter estimation and inference, while requires large amount of data to reliably learn the model with additional time dimensions. Therefore, it is hard to model temporal behavior in large scale real-world recommendation applications. In this work, we propose a new deep neural network based architecture that models the combination of user’s long term and short term temporal preferences. We also study the features that are effective for the recommendation applications. For instance, we use rich temporal features for a user from her search logs, and therefore to provide the context…


Link to Full Article: Multi-Rate Deep Learning for Temporal Recommendation